Admissibility of Recorded Dialogs
When it comes to legal implications and considerations, issues surrounding admissibility of voice recordings in court or at a tribunal can be complicated and complex. In general terms, a voice recording may be admissible if it satisfies certain conditions imposed by the Courts. The three most basic requirements for admissibility of evidence in this context are (1) authenticity, (2) source reliability, and (3) relevance.
The Court has discretion on how it may determine authenticity of the recording (see: R. v. Khelawon, 2006 SCC 57). The most basic requirement of the admissibility of this type of evidence is that a recording must be relevant. A recording is relevant if it is logically connected to a material fact in order to support a party’s case. The Court may consider the content, timing or circumstances of the recording to determine whether it is relevant to an issue between the parties.
In terms of authentication, the party must be able to establish that the recording was procedurally appropriate. In other words, the party who creates or controls the recording must be able to guarantee that the recording has not been altered or manipulated in any way. Similarly , the party who presents the voice recording in Court must attorn to its truthfulness and/or credibility.
Finally, source reliability is required for voice recording evidence to be admitted into Court. In other words, the Court must be satisfied that the voice is that of a party to the action or at least relevant to be able to draw an inference as to whom the speaker is.
Furthermore, other evidentiary rules may also affect whether or not a voice recording can be admitted in Court. Other rules about hearsay, witness testimony, the rule against opinion evidence (as it relates to expert witnesses) and/or the best evidence rule are all factors to consider when dealing with evidentiary issues involving the admissibility of voice recordings in Court.
The significance of voice recording evidence cannot be understated. The weight of such evidence can be determinative of whether or not a party is successful in litigation. For this reason, it is paramount to take steps to ensure that any voice recordings are accessible and admissible as evidence (if necessary) in order to strengthen your case.
Consent and Privacy Protections
Consent and privacy laws impact the admissibility of voice recordings in legal proceedings. In short, if you are in a one-party consent state, then as long as at least one person consents to the recording being made, it will be admissible in court. In a two-party consent state, all parties must consent for the conversation to be admissible.
Alabama, Michigan and Maryland do not have wiretap laws, so wiretap and eavesdropping laws are not applicable. As such, wiretaps, telephone taps or eavesdropping devices are generally admissible.
In the District of Columbia, only one party needs to consent.
Florida’s one-party consent law applies to "any wire or oral communication." The wiretap law does not apply to radio communications. In Florida, anything that is not a private communications service is not protected by the consent statute.
Georgia prohibits interception or disclosure of conversations without all parties’ consent.
Illinois prohibits the disclosure or use of a private conversation, meaning a conversation where someone reasonably expects the communication is private and would be extraordinarily intrusive to obtain. Exceptions apply. Courts weigh the balance between individual privacy and the need for law enforcement.
As of August 1, 2018, Louisiana is a two-party consent jurisdiction. Exceptions apply.
The Montana Statute against electronic eavesdropping only applies to "oral communications."
New Jersey has an exception for New Jersey. That said, Pennsylvania has an exception for courts in Pennsylvania and a protection for one-party consent circumstances.
Texas courts have to distinguish videorecordings from oral communications to apply the wiretap law.
Permissible Exceptions to Inadmissible Recordings
Though the general rule is that recorders are illegal if it is done without the consent of both parties, there are certain limited circumstances under which it will still be acceptable. These exceptions might include:
Public interest Where a reasonable person would be justified in believing that there was a public interest or safety in recording the conversation (as opposed to only the justice of the speaker). This should not be mistaken with the desired outcome of the evidence, but the fair and reasonable expectation of the background of the recording.
Prevention of serious crime People often think that prevention of serious crime ought to take precedence over the privacy of individuals. Courts have been reluctant to recognize this as a ground to convict someone where the recording was made illegally. There must be an immediate danger to the public- the recording cannot be justified on the premise that it might help to catch a criminal, or scare them off.
If the contents of the call are already known When a recording of the conversation contains information that is already known to the listener, the courts have been more willing to permit its introduction. Courts will sometimes rule that the contents of the call are somehow of a "public domain". Similarly, if information from a private conversation has been shared with others (and made public), then the information may no longer be as private.
Consent If a recording was made with the consent of one party (such as via a phone call with a disclaimer at the beginning letting the caller know the conversation is being recorded), then the recording may be more readily admissible.
Best Practices for Presenting a File as Proof
To help ensure the voice recordings you are using are admissible in court, here are some best practices to keep in mind: Store your recorded message on a secure server. If you’re using the recordings for a workplace harassment or sexual misconduct matter, be sure the server you use is secure and hosted in the United States, as the storage of the recording could become relevant in court. Any changes to the original recording should be documented. Recordings can suddenly become evidence when a surprise plaintiff shows up in court with an audio recording of what you said, so be sure that if you use editing software to delete or mask any portions of the recording, you maintain documentation of the changes. Get a sign-off on the recording. If you use recording devices in a workplace setting , notify your employees in advance of the use of such devices. Women are encouraged to avoid settings in which a dialogue is taking place – even in public – unless lights are on and accompanied by another person. Document the context of the recording and establish its authenticity. In explaining the material and history of the recording to the judge and jury, be as accurate and detailed as possible. Show them where the recording came from and how it was obtained. The more accurate the information you provide, the easier it will be to admit the recording into evidence. Ensure that the recorded voice is that of the person you claim it is. Even if you have evidence that it is that person’s voice, you may be asked to provide additional support, such as laying out additional facts showing it is that person. For example, you may need to show that the person had exclusive knowledge of the matter being discussed or that it was a picture of their person or property being discussed in the recording to establish that it was indeed that person’s voice.
Recordings in Key Court Cases
While the use of these recordings can sometimes give rise to privacy concerns, the reality is that they are often very helpful to the primary purpose for which they were created: establishing proof.
- In the case of LaLiberte v. Pac. Admiral Ins. Co., the United States Court of Appeals for the Ninth Circuit found that an auto insurance company did not have a right to call a consumer and record the call to solicit the renewal of their policy. The plaintiff did not give express consent for the insurer and Telemedia, the telemarketer, to record the call, and the Court found that a form of consent was not sufficient to overcome California’s law prohibiting the recording of telephone calls without consent.
- In United States v. Ziegler, 474 F.3d 1184 (9th Cir. 2007), the Court of Appeals for the Ninth Circuit affirmed a conviction based on voice recordings made by a confidential informant. These recordings contained information that was used to prove the elements of the crimes of conspiracy to manufacture methamphetamine and possession of an unregistered firearm. The Court examined the hearsay exception that allows admission of recorded statements if made by a coconspirator during the course of the conspiracy and in furtherance of the conspiracy. Statements fall within this hearsay exception if they are in furtherance of the conspiracy and were made during its pendency.
- In United States v. Davis, 592 F.3d 952 (8th Cir.2010), the Eighth Circuit affirmed the conviction of two men for drug crimes. An accomplice had provided a voice recording of the conversation between himself and the conviction, and the Court found that the evidence showed that the parties to the conversation were Davis and Romero. It was not necessary for the government witness to identify Davis’ voice, as the recipients of the conversation did not need to be identified.
Technology’s Impact on Admissibility
Since the advent of sound recording technology, courts and counsel have grappled with the admissibility of recorded sound as evidence. Audio recording technology has progressed rapidly in recent years, and some of the modern technology – the availability of inexpensive digital audio recording devices, for instance – stands to render the authentication of audio recordings more difficult and complex.
Authentication is a preliminary step that asks whether the recording is what it purports to be. The evidence proponent in a simple civil case must show that the recording is authentic but does not have the burden to show a specific speaker. The legal standard to authenticate a recording is a "reasonable certainty" that the recording is what it appears to be.
Consider a simple and common example: Susan and Tom hear John loudly yelling at their neighbor. Tom records the incident on his phone, and a few months later John sues the couple for defamation, arguing they were making a false statement about him when they made the recording. In this circumstance, it seems straightforward that the recording is, in fact, a recording of John yelling at the neighbor and therefore meets the reasonable certainty standard. But authentication is more complicated when there are no witnesses to the event.
Consider, then, two new methods that could be used to deceive courts and juries: voice duplication software and audio editing software. Voice duplication software can take real voices and, through a process of deep learning, emulate a person’s voice and correctly match their lips, head movements, and other gestures. Deep learning is a concept from artificial intelligence using neural networks, which are intended to mimic the way that the human brain works. The result is (potentially) perfect imitation, and it’s free and open source. For example, there is a website called Lyrebird, where users can upload 30 minutes of their own voice samples, and then download a new audio file that contains their exact voice saying whatever they inputted into Lyrebird’s text box. The output is fairly realistic, as long as you don’t pay attention to lip movements of the person in the video.
Likewise, audio editing software (e.g., Adobe Audition, GoldWave, Audacity) can also be used to edit out an unwanted word or phrase, remove background noises, or even change characteristics of a voice recording such as pitch or resonance. In practice, you might use audio editing software when you want to remove all of the "ums" and "likes" from a witness statement or to remove unwanted background noise .
The impact of these two technologies is clear: both raise the stakes for authentication. The Lyrebird software makes the "reasonable certainty" standard a difficult one to meet because the user may be able to generate realistic voice recordings of anyone with 30 minutes of audio. A voice identified by a particular individual as that of his or her own voice may not be genuine.
The obvious counter argument is that courts and juries are generally going to have a good faith belief that an audio recording is genuine unless and until the party opposing admissibility can tell the court why it’s not. This is, of course, true, but it’s possible that the sheer number of these realistic voice recordings will overwhelm judges and juries and cause them to simply accept the authenticity of any audio recording.
Audio editing software, on the other hand, raises authentication concerns only if it’s used to alter a recording that is being offered for authentication. If a recording is altered to remove an "um" or a background sound, the audio may still be authentic. However, the more serious issue is that when recordings are created at the time of the event, as they often will be with the use of cell phones and other audio recording devices, the alterations may not show up on audio editing software. This should cause courts to rethink how to get a handle on the admissibility of original recordings.
Each state also has various hearsay exceptions that might make the parties’ arguments as to admissibility moot in many cases. In California, for example, an exception exists for statements against interest, which by definition are statements that are harmful to the speaker and, therefore, not self-serving. The danger of deception is greatly reduced here.
An area that is gaining traction is forensics, which can complement the traditional "reasonable certainty" standard. Forensic analysts are tasked with analyzing different types of audio files, and when they do so, they look for sound waves or patterns that help identify different features of an audio file, such as background noise or the pitch of the human voice. Forensic scientists can also create spectrograms – x-ray-like images that highlight the patterns in the waveform – and then use those to show the authenticity of an audio recording or the presence of a person’s voice in a recording.
Unfortunately, the forensic analysis of voice recordings is in its infancy and the underlying science has not been well established. Nevertheless, reliance on forensic evidence is increasing – even in cases of deep fakes – and will certainly affect the admissibility of audio recordings.